Nuts & Volts - September 2006

following competition (that’s another story), when Tom commented that there were no good, inexpensive (I think he used the word cheap) speed controllers to use with these motors. So, as a group, we pointed at Ken and insisted that he design one. Since he also bought some silver bombers (five, I think), it was hard for him to refuse the challenge.

I’ll let Ken describe the design process ...

■ PHOTO 4. Parts salvaged from a Bomber.

KEN STARTS THE DESIGN PROCESS

I (Ken) bought my Silver Bombers with the intention of either using just the motors or turning them into some kind of tracked robot. A PWM driven controller is a must to build something like this.

There were no markings on the unit, so the only way to determine what kind of current it needed was to measure it. Using a current limited power supply, I ran it at six volts and determined that the current draw was about 2.5 amps with no load. I measured the winding resistance by carefully rotating the commutator until the lowest possible reading was obtained. This value would give a good indication of what the stall current would be. I measured about .715 ohms.

E=I*R give us 24V/.715 ohms or 33. 5 amps. I’ll take that as 35 amps for the design. With this in mind, we needed a controller that could safely withstand a 35 amp stall current. I asked some other club members to test their motors and they got similar results.

In the interest of full disclosure, I had already started a project like this some time ago, but was going for something much more sophisticated. I wanted a microcontroller driven unit with full quadrature or servo input that was capable of running some fairly sophisticated PID loops. Knowing this group as I do, I decided we needed something a little simpler to start with — a study H-bridge with simple control and as much protection built in as we could afford.

To give you a little background on me, I am a hardware/software designer with more than 35 years experience in many design venues.

Robots have been in my blood since I built my first one at age 12. Designing electronics and software is second nature to me, but some of the intrica-cies of machining and CNC were a mystery. That’s why I joined the group — to trade talents and help others out if I could. Phil and I regularly collaborate on various robotics projects.

A good design always starts with research. There is no point re-invent-ing the wheel, and lessons learned by the efforts of others are invaluable in producing a good, repeatable design.

I searched the web and looked at many robotic H-bridge projects. Most had some aspects of what I was looking for, but none quite put it all together in the professional (and sometimes quirky) manner I’m accustomed to.

For the purposes of this article, I’ll dispense with all the design equations needed to select parts and properly control the MOSFETs. These are readily available in many places; a quick search will ferret them out if you are really interested in this aspect of the design.

I settled on using the Intersil HIP4081A driver — a common and hardy solution with more than enough drive for a 50A controller. I had talked to the Intersil rep and he assured me it was a solid part and that they intended to support it for some time to come.

Coincidentally, I discovered that this is the same part used in the OSMC controller project and many of the humanoid robot designs in Japan. Couple this to some low Rds on FETs and you have the makings of a reliable driver.

The club members all have different controllers they are experimenting with. Some are using Bascom, others using embedded ‘C’ code. To accommodate this variety, I chose to provide the following inputs that can be driven by any microprocessor:

3. A PWM or brake input. When held low, motor braking action is initiated.

4. A fan enable pin. The fan can be turned off to conserve power when idle. The fan can also be PWM driven to provide fan speed control.

5. An auxiliary relay drive. Can be used to totally disconnect the motors or control brakes if using wheelchair motors.

6. A scaled voltage output for measuring the input voltage.

7. A heatsink temperature output scaled 10 mv/deg C.

All of the inputs and outputs are isolated (most optically), except the scaled input voltage. Fully isolating this would have made the project complete, but added too much cost.